Disclosed herein is a computer implemented method of determining a similarity score of a plurality of data records with a target data record in a data set. The similarity score allows a user to identify secondary data records, or pairs of data records, that disclose similar concepts. Also disclosed herein is a computer implemented method for presentation and visual navigation of a data set including related data records. The presentation of a data set using the disclosed method allows users of to quickly identify data records that are very similar to a data record of interest. The data set may include directly connected or indirectly connected patents.
|
1. A computer implemented method of determining a similarity score between two data records in a data record set, each data record including a predetermined data record strength value representative of the importance of each respective data record in the data record set, the method comprising:
identifying a target data record from the data record set;
identifying data records within the data set having primary and secondary connections with the target record to form a data record map, such records being primary data records having a direct connection with the target record and secondary data records being data records having a direct connection with a primary data record of the target record, wherein each primary and secondary data record is distinguished from the target record;
for at least one of the secondary data records, determining an importance value and a relevance value,
wherein the importance value of each secondary data record is determined based on:
interconnections between the secondary data record and other data records in the data record map; and
the predetermined data record strength value of each of the other data records having an interconnection with the secondary data record, and
wherein the relevance value of each secondary data record is determined based on:
interconnections between the secondary data records and other data records in the data record map relative to the target record; and
combining the importance value and the relevance value to provide the similarity score between the secondary data record and the target data record, wherein the similarity score allows a user to identify secondary data records that disclose similar concepts to the target data record.
2. A computer implemented method in accordance with
calculating a connection strength between the secondary data record and each primary data record connected to the secondary data record in the data map;
determining the predetermined data record strength value of each of the connected primary data records;
calculating a connection strength between the secondary data record with other secondary data records in the data map; and
determining the data record strength value of each of the secondary data records in the data set;
wherein the connection strengths and predetermined data record strength values are combined to form the importance value.
3. A computer implemented method in accordance with
counting the number of times the secondary data record is connected to primary data records in the data map to form a first vote; and
counting the number of times the secondary data record is connected another secondary data record in the data map to form a second vote;
wherein the first and second votes are combined to form the relevance value.
4. A computer implemented method in accordance with
calculating a connection distance between the secondary data records and the target data record in dependence on the connection strengths.
5. A computer implemented method in accordance with
6. A computer implemented method in accordance with
7. A computer implemented method in accordance with
8. A computer implemented method in accordance with
9. A computer implemented method in accordance with
10. A computer implemented method according to
|
|||||||||||||||||||||||||||||
This application is a Continuation of International Application No. PCT/AU2013/001223, International Filing Date Oct. 18, 2013, and which claims the benefit of Provisional Patent Application No. 61/795,579, filed Oct. 19, 2012, the disclosures of both applications being incorporated herein by reference.
The present invention relates to a system and method for visually navigating data sets including one or more networks of related data records, and particularly, although not exclusively to a system and method for interactive visual searching of intellectual property data sets, such as patent databases.
With the advent of cheap and powerful computing systems and the development of the electronic database, there has been an explosion in the collection and electronic storage of data related to almost all areas of technology, industry, commerce and society. Data is generally held, in many instances, in the form of a “record”, which typically comprises a series of attributes that describe a real world object or event. For example, one type of data record is a health record, which holds information regarding the attributes of a given person, such as their height, gender, weight, existing and past medical conditions, treatments undertaken etc. Another type of data record is that describing a scientific publication wherein a plurality of such data records may form a set and be held for example in a database of publications. Such a publications database can include attributes regarding the publications, such as the authors of each publication, citations or references to other publications, publication date and the subject matter of each publication.
Another structured set of data is data describing intellectual property rights, such as patent data records or trade mark data records. Many countries have legal regimes where owners or creators of intellectual property can register their rights to an invention, a sign and/or a design. Such records are highly structured and include a large number of attributes, such as a date of filing, the name of the owner or applicant, the names of the inventors or authors, data regarding the history of the invention and particular intellectual property office classification codes, such as the IPC (International Patent Classification) code, plus other attributes that describe the nature of the intellectual property right.
As patent data is effectively a record of innovative activity, value can be derived from searching patent data to extract commercially useful information. However, as an ever growing number of patents are filed every year, due to a constant increase in the rate of technological development and a greater awareness of the legal rights covering inventions, patent databases now contain millions or tens of millions of records, and in turn each patent data record contains a large and complex set of attributes. Therefore, traditional methods for searching such databases (such as by looking for keywords in the title, abstract or applicant details attributes) can lack precision, are prone to error and can return large and unwieldy data sets.
One method for selecting, analysing and visualising related database records utilises the network paradigm in view of the relationships that exist between and amongst at least some of the records. US Publication 2010/0106752 (Eckardt, III et al.) for example describes a network visualisation system and method for making sense of sets of related database records or documents by providing a network graphical representation of the records. However the difficulties inherent in analysing and graphically representing large and complex data sets, such as the representation of more than 1000 patent documents pictured in FIG. 13 of the '752 publication, are recognised. Eckardt considers at par [0177] that it is difficult to determine what is to be understood from this network graph of patent documents in which the nodes represent documents and the links are citation linkages.
Furthermore, without seeking professional assistance and studying each patent specification in detail, it is difficult to judge the relative worth or “merit” of a particular patent, or the underlying invention protected by the patent, in comparison to other patents and patented inventions. As such, traditional search methodologies struggle to adequately provide any sophisticated or high level information regarding the relative merit or worth of a patent.
In one proposal, U.S. Pat. No. 7,716,226 (Barney) describes a method and system for probabilistically quantifying and visualizing relevance between two or more citationally or contextually related data objects, in the context of statistically rating, valuing and analysing intellectual property assets including patents, patent applications and related documents. However, Barney relies on probabilistic analysis of patent documents particularly utilizing a multi-variate regression to provide a visual map. This approach has inherent drawbacks including inherent inaccuracies associated with averaging used in probabilistic techniques.
Disclosed herein is a computer implemented method of determining a similarity score of a plurality of data records with a target data record. The method comprises the steps of:
In some forms, the importance value for each of the at least one secondary data records is determined by;
In some forms, the relevance value for each of the at least one secondary data records is determined by;
In some forms, the relevance value for each of the at least one secondary data records is determined by;
In some forms, the target data record is identified by a user. In at least one embodiment, the importance value and the relevance value are attributed a weighting and combined to calculate the similarity score.
In some forms, the data records are patents, the primary data record being connected to the target data record in that they share a citation. In some forms, the secondary data records are connected to the primary data records in that they share a citation. In at least one embodiment, the data map shows connections between data records.
Also disclosed herein is a system for ranking, reducing and presenting data records. The system may comprise:
In a second aspect, the present disclosure provides a computer implemented method for presentation and visual navigation of a data set including related data records, the method including the steps of:
In an embodiment, the representation of each node reflects a score for the data record determined from the degree of connectivity with all other data records in the data set. Preferably, upon selection of a display icon, a visual cue is provided in the visual representation to guide attention to a node representing the next most relevant associated data record. Suitably the next most relevant data record is visually cued, at least in part, on the basis of the score determined for each of the associated data records. Most suitably, the associated data record with the next highest score is cued for user review.
In another embodiment, each link reflects the strength of the association between respective interconnected data records. Preferably the width and/or length of the links are proportional to the number of related data records associated with each pair of linked data records.
In a third aspect, the present disclosure provides a computer implemented method for presentation and visual navigation of a data set including related data records, the method including the steps of:
wherein:
Presenting phantom nodes, in the form of second order patents, as part of the network visual representation provides a number of advantages to the user (for example a patent examiner) of the computer implemented method. Including second order patents allows a patent examiner to quickly and efficiently identify data records that do not have a direct citation, i.e. patents that were not cited by examiners or patent applicants, with a patent of interest. This allows users of the method to quickly identify patents that are very similar to the patent of interest. Further, presenting patents in this way allows companies to easily find other companies that are operating in a very similar technical field.
In an embodiment, the score allocated to each data record has a value calculated from the number and degree of connections to other data records. Suitably the score may be normalised over all the records in the data set. In a particular embodiment wherein the data set is a patent database, the score allocated to a patent data record is calculated by a weighted combination of number of forward citations, number of backward citations and the relative age of the citations. Suitably the score allocated to each patent data record is normalised against the average score of patents in the patent database less than a predetermined period, for example 20 years.
In a forth aspect, the present disclosure provides a computer implemented method for presentation and visual navigation of a data set including related data records, the method including the steps of:
In an embodiment, the separate zone includes a list of the attribute information for each copied data record. The displaying step may further include an active window facilitating user entry of comments and/or assignment of a ranking to the data record represented by the selected node. Preferably, the user comments and/or the assigned ranking may be copied together with the attribute information, thereby providing an audit trail for viewing, storage or output, for example by printing.
In a fifth aspect, the present disclosure relates to a system for presentation and visual navigation of a data set including related data records, the system comprising:
a processor arranged for access to a data set including a plurality of related records;
the processor associated with an interface and further arranged to:
repeating the processor implemented steps utilising an identifier of the designated data record.
In a sixth aspect, the present disclosure relates to a system for presentation and visual navigation of a data set including related data records, the system comprising:
a processor arranged for access to a data set including a plurality of related records;
the processor associated with an interface and further arranged to:
wherein:
In a seventh aspect, the present disclosure relates to a system for presentation and visual navigation of a data set including related data records, the system comprising:
a processor arranged for access to a data set including a plurality of related records;
the processor associated with an interface and further arranged to:
In an eighth aspect, the present disclosure provides computer readable media containing sequences of instructions which, when executed by one or more processors, executes the steps of a method in accordance with any one of the second to forth, or tenth aspects.
In a ninth aspect, the present disclosure provides transmission or reception of a computer data signal comprising at least one encoded sequence of instructions from the eighth aspect.
In a tenth aspect, a computer implemented method of determining a similarity score of at least one pair of data records is disclosed. The method comprising the steps of:
In some forms, the method of determining a similarity score of at least one pair of data records includes the step of:
In some forms, the method of determining a similarity score of at least one pair of data records includes the step of:
In some forms, the connection strength is calculated by:
In some forms, the method of determining a similarity score of at least one pair of data records includes the step of:
In some forms, the data records are patents, the patents being directly connected in that they share a citation.
Also disclosed herein is a system for ranking, reducing and presenting at least one pair of data records, the system comprising:
In some forms, the system further comprises an attribute filter, the attribute filter operable to remove data records from the presentation.
In some forms, reduction of the data records caused by operation of the attribute filter does not require of regeneration of the presentation.
In some forms, reduction of the data records caused by operation of the attribute filter temporarily removes the data records from the presentation.
In some forms, the attribute filter also removes links associated with the removed data record to further reduce clutter from the presentation.
In some forms, the attribute filter is a percentage slider operable by a user to remove the data records.
Notwithstanding any other embodiments that may fall within the scope of the present disclosure, several embodiments of the present invention will now be described, by way of example, with reference to the accompanying drawings in which:
Referring to
In this embodiment, there is provided a local database containing a set of records, such as patent data records. In the embodiment described herein, the records may have been selected and collated in accordance with a co-pending application filed by the applicant, entitled “A system, method and computer program for preparing data for analysis”, published as US 2012/0011132, which is herein incorporated by reference. In another embodiment, it will be understood that the system may access a separately located and/or administered database containing patent data records. The database may be separately administered by a Government authority or third party.
Referring to
The server 100 may include storage devices such as a disk drive 108 which may encompass solid state drives, hard disk drives, optical drives or magnetic tape drives. The server 100 may also use a single disk drive or multiple disk drives. The server 100 may also have a suitable operating system 130 which resides on the disk drive 108 or in the ROM of the server 100. The system has a database 120 residing on a disk or other storage device which is arranged to store at least one record 122 providing, in at least one embodiment, a plurality of records containing patent related data. The database 120 is in communication with an interface (comprising one or more of the abovementioned components), which is implemented by computer software residing on the system 100.
An interface 202 provides a facility by which a user may input commands, instructions or requests to the server 100 for execution or processing, including an arrangement of hardware devices and software functionality. The interface is in connection with the processor and is arranged to facilitate both the entering of user commands and the display of information to the user. The interface 202 may be implemented with input devices such as keyboards 116, touch-pads, a mouse 118 or other pointing devices and, in another example embodiment, the interface 202 includes software modules which may be arranged to receive inputs, requests or data through a network connection, including Ethernet, Wi-Fi, Fire-Wire, USB or the like. The interface 202 of the present embodiment provides for the appropriate visualisation of patent related data, including before, during and/or after substantive analysis.
The following description provides a series of visualisation techniques utilised by the system to present, navigate and interactively analyse large data sets. While the techniques are described as discrete components, it will be understood that the techniques may be used in conjunction to provide a rich and deep visualisation of any relevant data set. The system and associated software application allow the user to apply the techniques in any suitable sequence, to provide the user with the most appropriate visualisation for their particular requirements.
Before describing the techniques in detail, it is instructive to provide some definitions, so that the reader may better understand the background to these techniques.
First, the techniques which are described herein apply principally to “network” visualisations. A network visualisation is a visual map or diagram which displays a plurality of entities and the relationship between the entities. For example, a map of train stations is a good illustration of a simple network map. Each dot or “node” represents a train station and each line or “link” represents a rail connection between the two stations. By using simple graphical devices such as nodes and links, a large amount of information about an entire rail network can be conveyed in a small space, and more importantly, in a manner that is intuitively understandable to a user (e.g. a train commuter). The map/diagram may be provided in a two dimensional format, or a three dimensional format, depending on the relative complexity of the data that is being displayed.
Secondly, in the context of the embodiment described herein, patent data is visualised using a network visualisation technique. Each node represents a patent or patent application, and each link represents a common shared attribute value, such as a common citation (i.e. a backward citation or forward citation).
Lastly, it will be understood that any suitable visualisation techniques/software components may be utilised to create a visual image of the network.
With these points in mind, the various techniques utilised by the system and method, which may be implemented at least in part as a software application, an embodiment of which will now be described with reference to
The main software module executes a method having a process flow 210 as depicted in
In step 218, a sub-module is called which passes the identifier for the purpose of determining data records in the set that are associated with the identified record and, as a consequence allocates, a score to each associated record (and the identified record) determined from the degree of connectivity with all other records in the data set. This will be further described in relation to
Upon conclusion of step 222, the process actively awaits such further input from the user. In the embodiment, “selection” of patent records at decision step 224 is effected by the user navigating through the network by moving from one node to another node using a pointing device, such as mouse 118, or by a pre-selected key stroke/s on keyboard 116. If required the framework can include a “Next” button 501, by which the user a guided to the patent record having the next highest “score.” This is helpful in approaching networks with relatively tightly grouped nodes, as will be further explained in relation to
The operation at decision step 226 concerns the designation of a new data record (as distinct from mere display of attributes) for re-generation of a fresh visual representation. This is effected in the embodiment and with reference to
We turn now to a discussion of the process relating to the sub-module for determining data records in the set that are associated with the identified record and, as a consequence allocates, a score to each associated record (and the identified record) determined from the degree of connectivity with all other records in the data set. More particularly we will discuss how the score may be deterministically calculated. In terms of the embodiment, the connecting attribute is citations and accordingly nodes are linked together in the representation on the basis of citation, preferably links have a first appearance for forward citations and another contrasting appearance for backward citations. Turning to the process depicted in
The subsequent steps involve the calculation of some intermediate metrics about the scale and depth of connections with other data records. Second step 232 requires the calculation for each data record of the number of first degree connections: Fc, conveniently referred to as “friends.” Third step 234 then requires the calculation for each pair of first degree connections (or each pair of friends) of the number of first degree connections that those friends have in common (“friends of friends”), summed across all pairs of friends SFc and conveniently referred to “total shared friends.” By way of example if a particular record had just 2 “friends” (i.e. first order citation links) and each of those friend 2 records themselves had 5 friend records in common, the total shared friends would calculate as: 2×5=10.
The next calculation step 236 then involves counting for each data record the number of forward links, i.e. to temporally later records, to provide a forward link count: FLc. Optionally, the forward link count is subject to a temporal weighting curve, which discounts the older forward citations. An example of weights for a suitable weighting curve is set out below in Table 1.
TABLE 1
Forward link discounts
1 yr
2 yrs
3 yrs
4 yrs
5 yrs
6 yrs
7 yrs
8 yrs
9 yrs
10 yrs
1.0
1.0
1.0.
1.0
0.84
0.7
0.6
0.52
0.46
0.4
11 yrs
12 yrs
13 yrs
14 yrs
15 yrs
16 yrs
17 yrs
18 yrs
19 yrs
20 yrs
0.36
0.32
0.286
0.26
0.24
0.22
0.2
0.18
0.156
0.144
Recent forward citations in a patent record example, generally suggest recent activity, which is good. But if a patent is 15 years old or more, and all of its forward citations were soon afterwards, this indicates that there is no recent related activity, and so such forward citations should be discounted heavily, such by this temporal weighting. A further adjustment, particularly for patents less than 5 years old, is to pro-rata the forward citations up—for example:
Pro-rata age corrected FLc=Age corrected FLc*5/3 (1)
Finally in step 238, the score is then the factored sum of two components SFc and FLc subsequent to normalisation against the average value for records in the data set, for example (in the case of patent records) patents issued in the past 20 years.
Score=f1*|SFc|+f2*|FLc| (2)
A main display zone 530 reserved for a visual representation of a network of data records is bordered to the left by a “Key” button 524 and by a “Tips” button 526, together with associated drop-down lists. Upon actuation of the Key button, the interface lists particular icon sizes and appearance codes, such as colour, applied to elements of the visual representation, and by actuation of the Tips button 526 a further drop-down list suggests actions currently available to a user. A user selected results zone 528 borders the right side of the main display zone 530, which in the present embodiment is arranged to be displayed in the form of a list. The list can include, for example, each data record's identifier together with other columns for attributes such as user relevance ranking, document title, owner, date and system calculated score (termed “global score” in a specific embodiment). The operation of the interface will now be described in relation to a particular example.
A first example of the operation of the presentation and visual navigation method and system of the first embodiment is now described in relation to
Each node is shown as a dot that represents an underlying patent record, with the dot 534 surrounded by a contrasting (such as coloured red) circle, here being the '968 patent, to indicate the node of the ‘focus’ patent of the patent network visual representation. All of the patents associated with the '968 patent are interconnected nodes of the network, wherein the nodes are represented by dots in a light grey colour with links represented by arcuate lines of contrasting appearance, such as colours. In some embodiments purple lines 536 are utilised to show backward citation links being a first sub-set of patent record associations, and green lines 538 are utilised to show forward citation links being a second sub-set of patent record associations. Suitably the size or area of the dots represent a score which is determined from the relative strength of interconnections to each patent, which score effectively indicates the influence (interconnectedness) of the respective patents in the network. In the present example, the '968 patent has a score of 0.24. This is a lower than average score in the embodiment, since the average interconnection strength of all US patents granted over the last 20 years is arranged in the determination to be unity (1), by normalising the results. If required, the thickness of lines may also be arranged to reflect the strength of connections between two particular nodes, such as by counting other data records that are associated in common with a pair of patent records and sometimes referred to as “shared friends”.
Upon initial generation of the patent network representation 532, a pop-up summary box 540 containing pre-selected attribute information of the patent, here “focus” patent '968, and the patent's score is presented. Here the attribute information is selected to include the patent title, number, year of grant, assignee/owner name, and (in a subsidiary box 542 showing) any image available from the patent record. An image is considered a particular powerful way for a user reviewing many records in quick succession to either quickly dismiss the record as “not relevant” or make a further inquiry into the details of the record (discussed in relation to
The patent summary box 540 further includes (at a top left corner) a rating button 550, which may be conveniently cycled, by user actuation, through a number of pre-set ratings. The user ratings applied to a patent record may also result in selective colour coding of the respective node, for example search relevance including: rating “0” for not relevant—with node colour green; rating “1” for potentially relevant—with node colour orange; and rating “2” for relevant—with node colour magenta. Reference is made to
Returning to
The visual representation 532 also shows nodes of the more highly scored patents that are connected to the patents that are themselves connected to the '968 patent. These indirectly connected patents could be regarded as influential ‘friends of friends’. In terms of the visual representation, the nodes are referred to as “phantom” nodes since they are presented at lower display intensity, and preferably feature transparency. Accordingly a phantom node, such as node 560, can easily be identified as the nodes wherein underlying links 562 are visible in
In contrast, as opposed to the full intensity nodes, such as node 564, wherein connection lines, such as link 538, are hidden behind the dots for the directly connected nodes, as shown in
The applicant has found that the patent records underlying phantom nodes, such as node 560, can be very valuable. Such patent records can contain potentially relevant disclosures that were not considered during official examination but still may be relevant to consideration of an invention being searched.
Turning now to
At step 810, the ‘importance value’ of the second order patents is calculated. The importance value takes into account the strength of the connection between a first order patent a second order patent, the strength of the connected first order patent, the strength of the connection between second order nodes in the data map (for example second order patents that share a citation but do not both share a citation with a first order patent) and the strength of this connected second order patent. The importance value gives a good indication of the ‘importantness’ of the second order patents, and is described in further detail by way of an example shown in
TABLE 2
Link Values between the focus patent and first order patents
Connection
Link Value
A-I
1
A-H
1
A-J
2
A-B
3
A-C
2
The strength of the first order patent that a second order patent is connected to may be a predetermined value, in the form of a ‘global score’, assigned to each patent in the data set before searching is performed. The global score may be related to the data set, not the data map, in that it is a predetermined value calculated for each patent in the data set before a focus patent is identified. Factors that may be included in the global score include a patents age, how many forwards citations it has and the citation rate over time. In some forms, the citation rate over time is normalised to allow relatively recent patents to have a high global score. For a network of patents, when a new patent is added to the data set, the global score of affected patents may be updated before a search is performed. The strength of connections between two second order patents may be calculated in much the same way that link values were calculated between first order patents and the focus patent. Table 3 details the link value for the second order patents in
TABLE 3
Link Values between first order patent and second order patents
Connection
Link Value
C-G
1
C-F
1
C-D
2
C-B
3
B-E
2
C-E
3
D-E
2
The strength of the second order patents, again the global score, is also taken into account when calculating the importance value of the second order patents. The four factors (strength of the first order patent, strength of the link between first and second order patents, strength of the second order patent and the strength of the link between second order patents) may be then given a weighting and an overall importance value is calculated for each of the second order patents. Table 4 details an example calculation of the importance value for second order patent E in
TABLE 4
Importance Value of E
Weight-
General Factors
Specific Factors
Value
ing
Score
Link value between first order
Link Value B-E
2
0.5
1
node and second order node
Link Value C-E
3
0.5
1.5
Global score of connected
Global score of B
3.2
0.25
0.8
first order node
Global score of C
0.32
0.25
0.08
Link value between second
Link Value D-E
2
0.15
0.30
order node and another second
order node
Global score of connected
Global score of D
6.1
0.1
0.61
second order node
Importance of
4.29
‘ghost patent’ E
It should be noted that the four factors used by way of example with reference to
At step 812, the relevance value is calculated for each of the second order patents. The relevance value gives a good indication of the relevancy of the second order patents to the focus patent within the data map. The relevancy value is calculated using a voting methodology, where each of the patents in the data map vote for each other. The number of times that each second order patent is referenced by a first order patent in the data map is calculated, and the reference is given two votes (the first order patent gives two votes to each second order patent it is connected to). The number of times that each second order patent is referenced by a second order patent in the data map is calculated, and the reference is given 1 vote (the second order patents give one vote to each second order patent they are directly connected to). In this example, the direction of the citation is taken into account. ‘By’ refers to a forward citation from one data record to another data record. For example, in relation to
TABLE 5
Relevancy Value for Patent E
Connection
Votes
C-E
2 votes
D-E
1 vote
B-E
2 votes
The relevancy value of patent E is therefore 5 (the summed votes from the connected first and second order patents). The overall similarity score of the second order patent, the ‘ghost patent’, may then be calculated by giving a weighting to the importance value and the relevancy value. In one form, the importance value is given a weighting of 0.25 and the relevancy value is given a weighting of 0.75. Usually the relevancy value is attributed greater significance than the importance value, as the relevancy value is more specific to the search enquiry. The score of ghost patent E in
As will be immediately apparent to the skilled addressee, there will usually be a much greater number of second order patents than first order patents. Presenting all of the second order patents may therefore clutter a presentation. One way to deal with this problem is to reduce the number of second order patents. The method of reducing includes the steps of determining the ideal number of ghost patents to be displayed (this is based on the number of first order patents in the network), which in turn determines the reduction severity; ranking the second ghost patents by distributing the similarity scores over percentile buckets (for example buckets of 0% increasing in 10% increments to 100% will give 10 buckets) and then determining the reduction point that gives less than the required number of second order patents. The percentile buckets ensure that two second order patents with the same score both survive the reduction. The reduced set of ghost patents may then be presented with the first order patents, as is shown in
A system, in the form of a computer, may be provided to rank, reduce and present data records. The system may comprise an identifying means, in the form of a user input into a computer program, for identifying the focus patent. The system may also comprise a calculating means, in the form of a processor, for calculating the similarity score of the secondary data records. The secondary data records may be ranked in dependence on the similarity score as described above. Further, a means, again in the form of processor, may be provided for reducing the secondary data records. Also, a display means, in the form of a computer screen, for presenting the reduced secondary data records with the primary and target data records may be provided. The presentation allows the user to identify primary and secondary data records that are directed toward a similar concept to the focus patent.
Returning to
One method of determining the connection distance between directly connected data records is to convert the connection strength into a value between zero and one, as detailed in Table 6 with reference to the connection strength (link values) determined in Tables 2 & 3. In this example, the conversion factor between link value and connection distance is 1/sqrt(connection strength). The conversion from connection strength to a connection distance can be a variety of forms. This could even be a sum of squares, or another form of conversion that has the same effect of using the connection strength to determine the similarity between two indirectly connected data records.
Using the conversion factor 1/sqrt(connection strength), a high link value becomes a small connection distance. A pair of indirectly connected data records may then be identified, where indirectly connected data records are connected via at least two pairs of directly connected data records. For example, data records I and D in
TABLE 6
Connection distance between pairs
of directly connected data records
Connection
Connection Strength
Connection Distance
A-I
1
1.00
A-H
1
1.00
A-J
2
0.71
A-B
3
0.58
A-C
2
0.71
C-G
1
1.00
C-F
1
1.00
C-D
2
0.71
C-B
3
0.58
B-E
2
0.71
C-E
3
0.58
D-E
2
0.71
B-J
2
0.71
The similarity score of the pair of directly and indirectly connected data records may then be calculated. This calculation is in dependence on the connection distance between the directly connected data records. For example, the connection distance between data records I and D could be the sum of the connection distances between each of the directly connected pairs, D-E, B-E, A-B and I-A, or, I-A, A-J, J-B, B-C, C-E and E-D. The shortest connection distance may be allocated to each indirectly connected pair of data records. One method to determine the shortest connection distance between two indirectly connected data records is to use the A* pathway algorithm, or any other suitable shortest pathway finding algorithm. For I and D, the shortest connection distance is via directly connected pairs I-A (1), A-C (0.71) and C-D (0.71), being 2.42. The similarity score for each pair of documents in the data map is calculated, the presentation of which allows a user to identify pairs of directly or indirectly connected data records that disclose similar concepts that may be in similar or dissimilar technical fields. For example, a patent in the technical field of telecommunicates may disclose or claim similar concepts to a patent in the technical field of computer science.
In the situation where there are a large number of data records, it may be beneficial to present only the most similar pairs of data records. The method may therefore also include the steps of setting a predetermined level for the similarity score between the pairs of indirectly connected data records, removing the pairs of indirectly connected data records that are below the predetermined level, and ranking the remaining pairs of indirectly connected data records in dependence on the similarity score. The ranked data records may be from the most similar to the least similar of the remaining pairs of data records. When the data records are in the form of patents, this ranking would provide the user with a list or graphic of the documents that detail the patents that are most likely to be directed towards a similar concept. Alternatively, the method may include the option of stopping once it identifies that the citation distance between any indirectly connected pair of data records is greater than a predetermined value. The ensuing report may consist of a listing of directly and indirectly connected pairs of data records, in the form of patent pairs, and their connection distance values, where the priority date may be used as a predictor to determine either patent anticipation or patent infringement.
For example, if Company A had a large patent portfolio and wanted to compare their portfolio with the large patent portfolio of Company B, this method would provide Company A with the ability to quickly assess their patents that are most similar to Company B's patents. This would be useful if Company A wanted to determine which of Company B's patents are most likely to anticipate or infringe their patents. Further, this method could also be used as an alternate to the method for determining the ‘relevancy value’, as described above, when identifying the most similar second order patents. In addition, this method could be used to determine data records that are a third or fourth order (or above) connections, which are very similar to a target data record.
Turning again to
In
The Focus button 556 provides another feature that can also assist in finding relevant prior art, namely the ability to ‘walk the network’, or refocus the patent network on another patent. In the present example, this has been done by actuating the Focus button in the patent summary box 540a for the '001, as patent discussed above. The result of the refocus, which effectively designates U.S. Pat. No. 6,349,001 as the identifier of the patent record of primary interest and re-generates another (new) visual representation 600 after determining which records in the patent data set are connected to the '001 patent and then “scoring” each connected patent, is shown in
It is apparent from
But are any of these patents in the re-generated representation 600 relevant to the '946 patent? Yes, some of them might be so further investigation by the user is merited. One way of “walking the network”, for example, is by the user actuating the “Next” button 501 in the bottom left corner of the framework. This employs the patent score value to select the next most relevant node 602, which relates to the patent record for U.S. Pat. No. 5,585,871—in relatively close proximity as depicted in
However, there are also quite a few phantom nodes for other patents, for example patent record U.S. Pat. No. 5,606,743 at node 604 in
Upon actuating the Next button 501, we are taken to node 706 which relates to patent record US51006179 which (from the “?” icon in the rating button) has not been previously reviewed/rated, as depicted in
In the discussion of the above example, we have shown how it is possible to find potentially relevant patent prior art, some of it missed by the patent examination processes, simply by starting with the patent number you are concerned with. This potentially relevant prior art could include:
directly connected patents,
ghost patents,
patents connected to directly connected patents (‘friends of friends’), or
patents connected to phantom patents (‘friends of friends of friends’).
A simple search using a conventional patent search for a patent which is close, but not close enough, can provide a suitable starting patent.
It is also worth considering what has not been done in this search example, viz:
We have not looked at any keywords or semantic terms, although this option is available if required. Different patent applicants can use different keywords for the same inventive concepts, and this can cause errors when searching for patent subject matter using keywords alone.
We have not looked at any patent classifications, such as the International Patent Classification (IPC) codes, which can be imprecise.
It has not been necessary have not spent hours and hours looking long lists of patents, many of them irrelevant. Instead we have relied on the power of citation networks to quickly identify relevant, some of which appear to be missed by the original patent examiner.
The network graph representation of the embodiment can provide a visual guiding mechanism based on node size, number and thickness of links, position of nodes in the representation with respect to the node for the initially searched “focus” patent (distance), and the position of nodes for patents with respect to other patents (clustering), including patents represented by phantom nodes. As will be apparent from the following discussion, users can intuitively select patents to look at based on one or more of the following considerations:
In another embodiment, the system and method may be arranged to provide a local importance score by comparing a global importance score (global score) to the same score for immediately connected (e.g. via citation linkage) patents (or and 2nd order, etc), rather than on keyword or IPC scores. Such a local importance score could be in comparison to the average of the same score of the connected patent, or a relative ranking in relation to these other scores, or any other suitable criteria.
In further embodiments of the invention, it is envisaged that the system will be able to capture a representation of the patent graph, rankings, etc. in the form of file accessed via a unique hyperlink, which can be emailed to a user's email address or any other desired/authorised party, and being able to reopen the visual presentation of the captured patent graph at the same point by selecting this hyperlink. The hyperlinked file is to includes all the users additional comments and annotations for the patent graph, and can be used to reopen the graph, and user rankings/comments etc from any suitably enabled browser.
In still further embodiments of the invention there may be provided certain patent database specific tools, including:
1) Family Member Searching
Imagine that you select an Australian patent number, and only find a few citation linkages. But there is a part of the patent summary box which allows the user to say open up a second box, which shows all known family members, including for example a US patent family member. By selecting this US family member (or any other family member), a new network is formed, based on this selected family member, but all search history and rankings is retained as with other patent refocussing.
Furthermore, the order of the family members in this second box could be ranked by their global score ratings.
2) Highlighting of Novelty/Obviousness Linkages
Some patent examiners prepare long lists of prior art patents, but only discuss a small number in their examination reports or office actions. In the US these references appear in a “Notice of References Cited” list, for example, since references are cited in support of novelty and/or obviousness rejections. In the later case, the order of the references may be important, as the first mentioned patent may be the strongest evidence against the patent application.
Where a list of such prior art patents, i.e. citation connections. Such connections could be a sub-set of the broader list of connections. A further level of detail could include highlighting such connections in the representations of patent data networks, with different highlighting for novelty objections, first obviousness objection, other obviousness objections. All of other connections could optionally be made to disappear, taking connected patents with them, so a tighter map is shown. The system could re-calculate a patent influence score based entirely on, or weighted towards, these novelty/inventive step connections. The system could also recalculate line thickness of links to be weighted towards these novelty/inventive step connections. As an alternative, we could provide a similar scheme based on X and Y weighting in say WO and/or EPO search reports.
3) Correction/Substitution of Patent Application Numbers
When you plug in a patent application or publication number in to the search field, and there is a granted patent with for the application with a different number, the number is translated and then shown on the screen as the equivalent granted number for that jurisdiction, not the application/publication number.
Alternative embodiments of a system and methods for analysing and visualising a for analysis are also described in co-pending U.S. Utility application Ser. No. 13/179,437 by the applicant entitled “A system, method and computer program for analysing and visualising data”, which is incorporated herein by reference. These embodiments are advantageous in that alternative schemes for analysis of data are provided that may be conveniently visualized using apparent variations of the present disclosure.
Although not required, the embodiments described with reference to the drawing figures can be implemented as an application programming interface (API) or as a series of libraries for use by a developer or can be included within another software application, such as a terminal or personal computer operating system or a portable computing device operating system. Generally, as program modules include routines, programs, objects, components and data files assisting in the performance of particular functions, the skilled person will understand that the functionality of the software application may be distributed across a number of routines, objects or components to achieve the same functionality desired herein.
It will also be appreciated that where the methods and systems of the present invention are either wholly implemented by computing system or partly implemented by computing systems then any appropriate computing system architecture may be utilised. This will include stand alone computers, network computers and dedicated hardware devices. Where the terms “computing system” and “computing device” are used, these terms are intended to cover any appropriate arrangement of computer hardware capable of implementing the function described.
It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. For example the system and method for visually navigating data sets including one or more networks of related data records of the invention may be adapted for utilisation with other connected data records, such as cross-referenced document collections, website pages, publications, trade mark records, court judgements, online objects, people networks, etc. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.
Any reference to prior art contained herein is not to be taken as an admission that the information is common general knowledge, unless otherwise indicated.
Although not required, embodiments described with reference to the drawings can be implemented to operate with any form of communication network operating with any type of communication protocol. Generally, where the underlying communication network or communication protocol includes additional routines, functionalities, infrastructure or packet formats, the skilled person will understand that the implementation of embodiments described including with reference to the drawings may be modified or optimized for operation with these additional routines, functionalities, infrastructure or packet formats.
Palmer, Ben, Spielthenner, Doris, Lloyd, Michael
| Patent | Priority | Assignee | Title |
| Patent | Priority | Assignee | Title |
| 5585871, | May 26 1995 | GOOGLE LLC | Multi-function display apparatus |
| 5606743, | Jan 22 1991 | Radio eyewear | |
| 6349001, | Oct 30 1997 | GOOGLE LLC | Eyeglass interface system |
| 7076485, | Mar 07 2001 | The MITRE Corporation | Method and system for finding similar records in mixed free-text and structured data |
| 7092857, | May 24 1999 | IPCentury AG | Neural network for computer-aided knowledge management |
| 7433884, | Sep 29 2004 | CHI Research, Inc.; Anthony F., Breitzman | Identification of licensing targets using citation neighbor search process |
| 7631968, | Nov 01 2006 | GOOGLE LLC | Cell phone display that clips onto eyeglasses |
| 7647287, | Nov 21 2008 | GOOGLE LLC | Suggesting a relationship for a node pair based upon shared connections versus total connections |
| 7716226, | Sep 27 2005 | PatentRatings, LLC | Method and system for probabilistically quantifying and visualizing relevance between two or more citationally or contextually related data objects |
| 7930301, | Mar 31 2003 | Microsoft Technology Licensing, LLC | System and method for searching computer files and returning identified files and associated files |
| 20040036716, | |||
| 20100106752, | |||
| 20120011132, | |||
| EP986789, | |||
| EP1184798, | |||
| JP2011138470, | |||
| WO72256, | |||
| WO2006001906, | |||
| WO2010065108, |
| Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
| Apr 17 2015 | Patent Analytics Holding Pty Ltd | (assignment on the face of the patent) | / | |||
| May 15 2015 | PALMER, BEN | Patent Analytics Holding Pty Ltd | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 035719 | /0099 | |
| May 15 2015 | SPIELTHENNER, DORIS | Patent Analytics Holding Pty Ltd | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 035719 | /0099 | |
| May 15 2015 | LLOYD, MICHAEL | Patent Analytics Holding Pty Ltd | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 035719 | /0099 |
| Date | Maintenance Fee Events |
| Sep 22 2020 | M2551: Payment of Maintenance Fee, 4th Yr, Small Entity. |
| Date | Maintenance Schedule |
| Sep 19 2020 | 4 years fee payment window open |
| Mar 19 2021 | 6 months grace period start (w surcharge) |
| Sep 19 2021 | patent expiry (for year 4) |
| Sep 19 2023 | 2 years to revive unintentionally abandoned end. (for year 4) |
| Sep 19 2024 | 8 years fee payment window open |
| Mar 19 2025 | 6 months grace period start (w surcharge) |
| Sep 19 2025 | patent expiry (for year 8) |
| Sep 19 2027 | 2 years to revive unintentionally abandoned end. (for year 8) |
| Sep 19 2028 | 12 years fee payment window open |
| Mar 19 2029 | 6 months grace period start (w surcharge) |
| Sep 19 2029 | patent expiry (for year 12) |
| Sep 19 2031 | 2 years to revive unintentionally abandoned end. (for year 12) |